Serum testosterone, testosterone replacement therapy and all-cause mortality in men with type 2 diabetes: retrospective consideration of the impact of PDE5 inhibitors and statins

SummaryBackground Low testosterone levels occur in over 40% of men with type 2 diabetes mellitus (T2DM) and have been associated with increased mortality. Testosterone replacement together with statins and phosphodiesterase 5 inhibitors (PDE5I) are widely used in men with T2DM. Purpose To determine the impact of testosterone and testosterone replacement therapy (TRT) on mortality and assess the independence of this effect by adjusting statistical models for statin and PDE5I use. Methods We studied 857 men with T2DM screened from five primary care practices during April 2007–April 2009. Of the 857 men, 175/637 men with serum total testosterone ≤ 12 nmol/l or free testosterone (FT) ≤ 0.25 nmol/l received TU for a mean of 3.8 ± 1.2 (SD) years. PDE5I and statins were prescribed to 175/857 and 662/857 men respectively. All-cause mortality was the primary end-point. Cox regression models were used to compare survival in the three testosterone level/treatment groups, the analysis adjusted for age, statin and PDE5I use, BMI, blood pressure and lipids. Results Compared with the Low T/untreated group, mortality in the Normal T/untreated (HR: 0.62, CI: 0.41–0.94) or Low T/treated (HR: 0.38, CI: 0.16–0.90) groups was significantly reduced. PDE5I use was significantly associated with reduced mortality (HR: 0.21, CI: 0.066–0.68). After repeating the Cox regression in the 682 men not given a PDE5I, mortality in the Normal T/untreated and Low T/treated groups was significantly lower than that in the reference Low T/untreated group. Mortality in the PDE5I/treated was significantly reduced compared with the PDE5I/untreated group (OR: 0.06, CI: 0.009–0.47). Conclusions Testosterone replacement therapy is independently associated with reduced mortality in men with T2DM. PDE5I use, included as a confounding factor, was associated with decreased mortality in all patients and, those not on TRT, suggesting independence of effect. The impact of PDE5I treatment on mortality (both HR and OR < 0.25) needs confirmation by independent studies.